How it works: WAVE calibration & machine learning

This is a brief overview of the tech supporting the WAVE's calibration process.

The WAVE uses a simple five-point calibration procedure during setup to map the IR reflections it reads to touch points on the edges and center of any screen on which it is set up. It dynamically detects the boundaries of the host screen during this process, allowing it to work with a variety of aspect ratios and screen sizes.

While you press on the points that appear on the screen, the WAVE learns not only where the edges of the tethered TV screen are, but which external light sources to ignore, and where to attenuate its signal sensitivity. This allows it to function in especially bright or dark environments, or on top of reflective surfaces, despite its dependence on reflected IR light to read touch inputs.

Though the calibration done during initial setup does not require the WAVE to connect to the internet, any subsequent calibrations once connected to Wi-Fi will make use of the WAVE's robust machine-learning library. The library is formed from the compiled user data of every successful calibration, which logs screen size, screen shape, and ambient lighting data along with the WAVE's serial number. Your WAVE uses this library to solve dead zone and sensitivity problems it may familiar with from similar WAVE calibrations in the past.

The result is each new WAVE setup progressively resolving sensitivity weaknesses across the whole network of in-use WAVE's.

Have more questions? Submit a request

Comments

Powered by Zendesk